Searching functionality for RAG

We are creating embeddings and storing them in pinecone, we will get one object from pinecone , that has embeddings, so we are thinking to store that object in AWS s3 and will perform searching, because we thought when we use pinecone for searching we need to create embeddings and store them every time while searching.we are using static data.
Is my thought is correct ?

Hey @anusha.gudipati. I think there’s some confusion here. Pinecone is a database, not an embedding service. You don’t have to recreate your embeddings every time you search against them, that’s what you’re using Pinecone for.

You do have to create an embedding for the search you’re running, to function as the query vector. But there’s no need to store your vectors in AWS or anywhere else. Just upsert them to Pinecone and they’re ready for you when you do your searches.

can you provide sample code, like searching questions from pinecone ?Assume like you already stored embeddings in index of pinecone.

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.